G.723.1 AUDIO ENCODER

Description

[0001] The present invention relates to low complexity encoders, and more particularly, to low complexity encoders for implementing recommendation G.723.1 of the International Telecommunication Union (ITU-T).

[0002] Lower complexity compressors/decompressors (codecs) may be preferred for some computationally intensive applications. If the complexity of the codec is the bottleneck in a system, complexity reduction is desirable and can result in a significant reduction in millions of instructions per second (MIPS) required to be executed by the encoder.

[0003] The ITU-T recommendation G.723.1, incorporated herein by reference, relates to dual rare speech coding for multimedia communications transmitting at 5.3 and 6.3 Kbps. The recommendation prescribes certain methods of implementation for each of these transmission rates. The 6.3 Kbps codec has better quality and uses Multi-Phase Maximum Likelihood Quantization (MP-MLQ) for fixed codebook excitation. The 5.3 Kbps codec uses Algebraic Code-Excited Linear Prediction (ACELP). A functional module of the codec which executes these two encoding methods bears almost half of the computational load of the entire G.723.1 speech coder. If the methods executed by the functional module are made to have a decreased computational load, the G.723.1 speech coder will have an increased efficiency.

[0004] Huijuan Cui et al.: "Audio as a support to low bit rate multimedia communication" International Conference on Communication Technology proceedings (IEEE Cat. No. 98EX243), ICCT'98, Beijing, china, 22-24 Oct. 1998, pages 544-547 discloses a dual rate encoding system a method for optimizing the original ITU-T G.723.1 speech codec. TO this aim, this document modifies the standard ITU-T G.723.1 speech codec as regards the MP-MLQ algorithm, by exploiting the fact that there is some correlation between the adjacent sub-frame pulse position patterns. Based on that, the fixed full codebook search is performed for the first and third frames only. The search for multi-pulse excitation of the second and fourth sub-frames is performed based on the previous sub-frame multi-pulse excitation result. Furthermore, only best gain levels are selected for different pitch lags.

[0005] US-A-5 717 825 teaches a modified ACELP algorithm by generally indicating the use of an algebraic codebook associated with focused search with adaptative threshold. However, this document does not teach how to compute the adaptative threshold for the codebook search.

[0006] The present invention provides a method of reducing the computational load of a dual rate encoding system according to claim 1. In particular, the encoding system is configured to transmit at a first transmission rate using a Multi-Pulse Maximum Likelihood Quantization (MP-MLQ) process or at a second transmission rate using an Algebraic Code-Excited Linear Prediction (ACELP) process, wherein the normal MP-MLQ process searches subframes of excitation signals according to a nominal number of gain scale factors in the execution of quantization steps for encoding the speech signals and the normal ACELP process imposes a first correlation threshold test for entering a last signal processing loop, the method including the step of:

for the MP-MLQ process, reducing the number of gain scale factors employed in the quantization steps, thereby reducing the number of gain searches, which in turn reduces the computational load; or

for the ACELP process, imposing a second correlation threshold test for entering a previous signal processing loop thereby reducing the number of times the previous signal processing loop and the last signal processing loop are entered, which in turn reduces the computational load.

[0007] The present invention further provides a dual rate speech coding system having a reduced computational load according to claim 11. In particular, the encoding system has Multi-Pulse Maximum Likelihood Quantization (MP-MLQ) processing means for transmitting at a first transmission rate and Algebraic Code-Excited Linear Prediction (ACELP) processing means for transmitting at a second transmission rate, wherein the normal MP-MLQ processing means searches subframes of excitation signals according to a nominal number of gain scale factors in quantization of the speech signals, and the normal ACELP processing means uses a first correlation threshold test for allowing entry into a last signal processing loop, wherein:

the MP-MLQ processing means has a reduced number of gain scale factors for reducing the number of gain searches and thereby reducing the computational load;

the ACELP processing means uses a second correlation threshold test for allowing entry into a previous signal processing loop which precedes said last signal loop entered in dependence on said first correlation threshold test, thereby reducing the number of times the previous signal processing loop and the last signal processing loop are entered, which in turn reduces the computational load.

[0008] Advantageously, embodiments of the invention simplify the ACELP and MP-MLQ methods by reducing the number of recursions which make less contribution to the metrics. This is achieved by selecting less gain levels or putting an extra threshold to decrease the chance to enter the most computational intensive loops.

[0009] Advantageously, the proposed encoder scheme is applicable for both ITU-T recommendations G.723.1 and G.723.1A. For ACELP excitation, further complexity reduction is possible by adjusting the thresholds. This complexity reduction for ACELP excitation is also applicable for G.729 and its annexes.

[0010] The present invention will now be described in further detail, by way of example only, with reference to the accompanying drawing.

[0011] Figure 1 is a block diagram of the G.723.1 speech coder.

[0012] Reference is also made to the following procedures which are appended to this description.

Procedure 1 is a pseudocode representation of the standard MP-MLQ procedure of the G.723.1 speech coder;

Procedure 2 is a pseudocode representation of the MP-MLQ procedure of an embodiment of the present invention;

Procedure 3 is a pseudocode representation of the standard ACELP procedure of the G.723.1 speech coder;

Procedure 4 is a pseudocode representation of the ACELP procedure of an embodiment of the present invention.

[0013] A MP-MLQ/ACELP block 10 for implementing the MP-MLQ and ACELP excitation methods is shown in Figure 1. These methods take up almost half of computational load of the whole codec. Since embodiments of the present invention only relate to these two fixed codebook excitation methods, the description relates only to these excitation techniques and not to other parts of the G.723.1 speech coder. Apart from the fixed codebook excitation part (i.e. block 10), all other modules are the same for the dual rate coders. The decoding scheme, for decoding bit streams encoded with the low complexity encoder, remains the same as for the normal ITU-T G.723.1 recommendation.

MP-MLQ Excitation (normal complexity)

[0014] The object of the quantization procedure is to find the optimized excitation e_u(n) which makes the mean square error minimum, based on an analysis by synthesis method. The excitation signal is given by

where G_u is the gain factor, δ(n) is a Dirac function, {α_k}_k=0..Np-1 and ξ_k are the signs (±1) and positions of the Dirac functions respectively, and N_p is the number of pulses, which is 5 for odd subframes and 6 for even subframes. The pulse positions in are either all odd or all even. This is indicated by a grid bit.

[0015] The scalar gain quantizer consists of 24 steps, of 3.2 dB each. Around the quantized value, G_u, additional gain values are selected within the range [G_u - 6.4dB; G_u+3.2dB]. The optimal combination of pulse locations and gains are then transmitted to the remaining encoder modules.

[0016] To improve the quality of speech with a short pitch period, the following additional procedure is used. If the pitch lag is less than 58 samples for a particular subframe, a train of Dirac functions with a period of the pitch index is used for each location ξ_k instead of a single Dirac function in the above quantization procedure. The choice between a train of Dirac functions or a single Dirac function to represent the residual signal is made based on the mean square error computation. The configuration which yields the lowest mean square error is selected.

[0017] Based on the above brief description of MP-MLQ, the optimization procedure is represented in pseudocode as shown in Procedure 1. The symbols InsCI inside the brackets are the cycles needed for a given processor; and the number of cycles if using, for example, a D950 processor. The D950 is a normal 16-bit fixed-point digital signal processor (DSP) made by STMicroelectronics. Other 16-bit fixed-point DSPs are the ADSP-2181 by Analog Devices and the TMS320C54x series by Texas Instruments. Although the number of instructions required to execute the same function may vary among different DSPs, the invention will still achieve a significant savings in MIPS for each appropriate DSP.

[0018] The worst case for MP-MLQ is that above optimization procedure is conducted twice when the pitch is less than 58 samples. The total number of cycles per subframe is given by

[0019] Therefore the total number of cycles per subframe for the procedure of Procedure 1 is 64368 if using the D950 processor.

ACELP excitation (normal complexity)

[0020] For the ACELP technique for fixed codebook excitation, a 17-bit algebraic codebook is used for the stochastic codebook excitation e_u2(n). Each fixed codcvector contains four non-zero pulses which can assume the signs and positions given in the following table.

Table 1

ACELP excitation codebook
Sign	Positions
±1	0	8	16	24	32	40	48	56
±1	2	10	18	26	34	42	50	58
±1	4	12	20	28	36	44	52	(60)
±1	6	14	22	30	38	46	54	(62)

[0021] In the table, all pulses are in the even positions but the positions of all pulses can be simultaneously shifted by one (to occupy odd positions) when requiring one extra bit. Note that the last position of each of the last two pulses falls outside the subframe boundary, which signifies that the pulse is not present. Each pulse position is encoded with 3 bits and each pulse sign is encoded with 1 bit. This gives a total of 16 bits for the 4 pulses. Further, an extra bit is used to encode the shift. The excitation sequence is defined as

   where ξ_k is the position of the k^th pulse and α_k is its sign (±1).

[0022] A focused search approach is used to simplify the search procedure. To limit the number of times entering the last loop, a threshold is applied and the last loop is entered only if this threshold is exceeded. The maximum number of times the loop can be entered is fixed so that a low percentage of the codebook is searched. The maximum absolute correlation C_max3 and the average correlation C_nv3 due to the contribution of the first three pulses are found prior to the codebook search. The threshold is given by:

[0023] The fourth loop is entered only if the absolute correlation (of the three pulses) exceeds thr3.
To further control the search, the number of times the last loop is entered (for the 4 subframes) is not allowed to exceed 600. (The average worst case per subframe is 150 times).

[0024] Based on the above brief description of ACELP, the optimization procedure is represented in pseudocode as shown in Procedure 3. In Procedure 3, InsCi is the number of instruction cycles, followed by an example number of cycles for the D950 implementation. The total cycles are calculated by

where time₃ is the number of times entering the last loop. At the worst case, the maximum number of time₃ is set to 150. Therefore the worst case cycles per 7.5 ms subframe are 62907 if using a D950 processor, which equates to 8.4 MIPS.

Lower Complexity Implementations

[0025] In embodiments of the invention the modules (codes) may be shared by both G.723.1 and the lower complexity implementation of the G.723.1 coder (LC-G.723.1). Preferably, the coding system is selectable between bit-exact G.723.1 and LC-G.723.1 coders, leading to an embedded system. This is shown by the procedure as follows:

[0026] For the low-complexity encoding of 6.3 Kbps and 5.3 Kbps codecs in accordance with the present invention, the operation procedures are shown in Procedure 2 and Procedure 4 respectively.

[0027] One of the characteristics of MP-MLQ is that the latter pulse contribution will be added upon the previous one and all pulses are scaled by one gain. For each new found pulse, the gain is further fine tuned within the range [-6.4dB;-3.2dB; 0; +3.2dB]. Since all pulses share one gain, the observation is that the gain level decreases as the number of found pulses increases. Due to the characteristic of MP-MLQ, the additional higher gain levels (0 and +3.2dB) are rarely selected. In this simplification, we only use two gain levels, i.e. -6.4 dB and -3.2 dB around the previous quantized gain. Therefore the number of instructions inside the gain searching loop can be decreased by about half for each subframe when the pitch lag is less than 58 samples.

[0028] The worst case number of cycles for MP-MLQ is calculated as:

[0029] For the D950 example, the total number of cycles per subframe is 39424.

[0030] For an adaptive codebook search, the worst case is when the pitch lag ≥ 58, which is just the opposite of fixed codebook excitation. If the number of gain levels decreases from 4 to 2 for fixed codebook excitation, the computational load is reduced from Equation (2) to Equation (6). To balance the computational load for all cases, the codes are also simplified for when the pitch lag ≥ 58. The number of searched gain levels is reduced from 4 to 3, i.e. -6.4, -3.2 and 0 dB. (please refer to Procedure 2).

[0031] The number of cycles per subframe for MP-MLQ with a pitch lag ≥ 58 is calculated as

[0032] The total number of cycles per subframe would then be 19826 for the D950 processor example.

[0033] Comparing Equations (2) and (6), 24944 cycles per subframe can be saved at worst case (of MP-MLQ) if using the D950 processor. This equates to a saving of 3.3 MIPS. For the normal case, in which MP-MLQ is conducted once, the saved cycles are 12358 per subframe, which equates to 1.65 MIPS. This unbalanced complexity reduction in the fixed codebook search (MP-MLQ) corresponds to the unbalanced computational load adaptive codebook search, in which, for example, about 30,000 and 46,000 cycles are needed respectively for the worst case and normal case of MP-MLQ.

[0034] A purpose of embodiments of the invention is to reduce the complexity for the worst case scenario (i.e., under the most intensive computational load). If the complexity is reduced in the worst case, the overall MIPS requirement is reduced accordingly. At the higher bit rate, the most complex modules are the fixed codebook excitation module (MP-MLQ) and adaptive excitation module. The complexity of these two modules changes depending on the pitch lag, while other modules are relatively stable in terms of computational load. Shown in Table 2 below is a comparison of the MIPS requirements for the worst case (pitch lag < 58 samples) and the normal case (pitch lag ≥58) for a D950 DSP.

Table 2

Complexity comparison 6.3kbits/s for one subframe (7.5ms)
6.3 kbit/s		Pitch lag ≥ 58			Pitch lag < 58 (Worst case)
		Adaptive	ML-LPQ	Sum	Adaptive	ML-LPQ Sum
Normal G.723.1	Cycles	46000	32184	78184	30000	64368	94368
	MIPS	6.13	4.29	10.42	4.0	8.58	12.58
LC G.723.1	Cycles	46000	19826	65826	30000	39424	69424
	MIPS	6.13	2.64	8.77	4.0	5.26	9.26

[0035] From Procedure 3 and Equation (5), it is apparent that any instructions inside the i₂ and i₃ loops will be executed hundreds of times. It may be advantageous to further limit the numbers entering these two loops. Instead of using one threshold, two thresholds are used. Both the maximum absolute correlation and the average correlation due to the contribution of the first two and three pulses, C_max2 and C_nv2, and C_max3 and C_nv3, are found prior to the codebook search. The thresholds are calculated by:

[0036] Now we have two thresholds. To further control the search, the average number of times the third and last loops are entered is not allowed to exceed 32 and 75 (for example), respectively for each subframe. The proposed low-complexity ACELP optimization procedure is modified as in Procedure 4.

[0037] The total number of cycles per subframe is given by:

where time₂ and time₃ are the number of times the processor enters into the 3^rd and 4^th loops respectively. For the worst case, the time₂ and time₃ are set to 32 and 75 respectively. Therefore the worst case number of cycles will become 36976. Comparing with Equation (5), 25932 cycles or 3.45 MIPS can be saved (if using the D950 processor).

[0038] It should be noted that further complexity reduction is simple to effect for this ACELP excitation by choosing smaller time₂ and time₃ parameters and corresponding higher thresholds. The proposed parameters for this LC-G.723.1 are based on the objective that LC-G.723.1 should have similar performances to G.723.1. If further reduction of complexity is needed, the performance will be smoothly degraded. For example, by increasing the threshold levels and corresponding allowed loop entry times time₂ and time₃ to 20 and 60 respectively, a further 1.01 MIPS can be saved.

Claims

1. A method of reducing the computational load of a dual rate encoding system, the encoding system being configured to transmit at a first transmission rate using a Multi-Pulse Maximum Likelihood Quantization (MP-MLQ) process or at a second transmission rate using an Algebraic Code-Excited Linear Prediction (ACELP) process, wherein the MP-MLQ process searches subframes of excitation signals according to a nominal number of gain scale factors in the execution of quantization steps for encoding the speech signals and the normal ACELP process imposes a first correlation threshold test for entering a lasst signal processing loop, characterized in that the method includes the step of:

for the MP-MLQ process, reducing the number of gain scale factors employed in the quantization steps, thereby reducing the number of gain searches, which in turn reduces the computational load; or

for the ACELP process, imposing a second correlation threshold test for entering a previous signal processing loop which precedes said last signal processing loop entered in dependence on said first correlation threshold test, thereby reducing the number of times the previous signal processing loop and the last signal processing loop are entered, which in turn reduces the computational load.

2. The method of claim 1, wherein the second threshold test is applicable for entry into the third of four signal processing loops.

3. The method of claim 2, wherein if the second transmission rate is applicable, the method further includes the step of substituting for the first threshold a higher threshold for entry into the fourth signal processing loop.

4. The method of claim 3, wherein if the second transmission rate is applicable, further including the step of limiting the number of times the third or fourth signal processing loop may be entered.

5. The method of claim 4, wherein the third or fourth signal processing loops may be entered up to 32 or 75 times respectively for each of the speech subframes.

6. The method of claim 5, wherein the dual rate coding system is generally in accordance with the ITU-T G.723.1 recommendation.

7. The method of claim 1, wherein if a pitch lag of the subframe is less than a predetermined parameter, the number of gain scale factors searched is reduced from four to two.

8. The method of claim 7, wherein if the pitch lag of the subframe is equal to or greater than the predetermined parameter, the number of gain scale factors searched is reduced from four to three.

9. The method of claim 8, wherein the predetermined parameter is 58.

10. The method of claim 1 or 9, wherein the quantization steps and a pre-search are executed once if the pitch lag is greater than or equal to 58 and twice if the pitch lag is less than 58.

11. A dual rate speech coding system having a reduced computational load, the encoding system having Multi-Pulse Maximum Likelihood Quantization (MP-MLQ) processing means for transmitting at a first transmission rate and Algebraic Code-Excited Linear Prediction (ACELP) processing means for transmitting at a second transmission rate, wherein the normal MP-MLQ processing means searches subframes of excitation signals according to a nominal number of gain scale factors in quantization of the speech signals, and the normal ACELP processing means uses a first correlation threshold test for allowing entry into a last signal processing loop, characterized in that:

the MP-MLQ processing means has a reduced number of gain scale factors for reducing the number of gain searches and thereby reducing the computational load;

the ACELP processing means uses a second correlation threshold test for allowing entry into a previous signal processing loop which precedes said last signal processing loop entered in dependence on said first correlation threshold test, thereby reducing the number of times the previous signal processing loop and the last signal processing loop are entered, which in turn reduces the computational load.

Ansprüche

1. Ein Verfahren zum Reduzieren der Rechenlast eines Kodiersystems mit zwei Raten bzw. Geschwindigkeiten, wobei das Kodiersystem konfiguriert ist, um mit einer ersten Übertragungsrate bzw. -Geschwindigkeit unter Verwendung eines Mehrfachimpuls-Maximimal-Wahrscheinlichkeits-Quantisierungs-(Multi-Pulse Maximimum Likelihood Quantization, MP-MLQ)-Prozesses oder mit einer zweiten Übertragungsrate unter Verwendung eines algebraischen kodeangeregten Linearprediktions-(Algebraic Code-Excited Linear Prediction, ACELP)-Prozesses zu senden, wobei der normale MP-MLQ-Prozess-Unterrahmen von Anregungssignalen absucht, und zwar gemäß einer nominalen Anzahl von Verstärkungsskalierungsfaktoren bei der Ausführung von Quantisierungsschritten bzw. Quantisierungsstufen zur Kodierung der Sprachsignale und der normale ACELP-Prozess macht einen ersten Korrelationschwellentest zum Aufrufen einer letzten Signalverarbeitungsschleife, dadurch gekennzeichnet, dass das Verfahren die folgenden Schritte aufweist:

für den MP-MLQ-Prozess Reduzieren der Anzahl der Verstärkungsskalierungsfaktoren, die bei den Quantisierungsschritten angewendet werden und reduziert dadurch die Anzahl der Verstärkungssuchen, was wiederum die Rechenlast reduziert; oder

für den ACELP-Prozess Einführen eines zweiten Korrelationsschwellentests zum Aufrufen einer vorhergehenden Signalverarbeitungsschleife, die der genannten letzten Signalverarbeitungsschleife vorangeht, die in Abhängigkeit von dem genannten ersten Korrelationsschwellentest aufgerufen wird und dadurch die Anzahl von Malen reduziert, die die vorhergehende Signalverarbeitungsschleife und die letzte Signalverarbeitungsschleife aufgerufen werden, was wiederum die Rechenlast reduziert.

2. Verfahren nach Anspruch 1, wobei der zweite Schwellentest anwendbar ist zum Eintritt in die dritte von vier Signalverarbeitungsschleifen.

3. Verfahren nach Anspruch 2, wobei, falls die zweite Übertragungsrate anwendbar ist, das Verfahren ferner den folgenden Schritt beinhaltet: Ersetzen der ersten Schwelle durch eine höhere Schwelle zum Eintritt in die vierte Signalverarbeitungsschleife.

4. Verfahren nach Anspruch 3, wobei, falls die zweite Übertragungsrate anwendbar ist, ferner den folgenden Schritt beinhaltet: Limitieren der Anzahl von Malen, die die dritte oder vierte Signalverarbeitungsschleife aufgerufen werden kann.

5. Verfahren nach Anspruch 4, wobei die dritten oder vierten Signalverarbeitungsschleifen bis zu 32 oder entsprechend 75 mal für jeden der Sprachunterrahmen aufgerufen werden können.

6. Verfahren nach Anspruch 5, wobei das Kodiersystem mit zwei Raten im allgemeinen in Übereinstimmung mit der ITU-T G.723.1 Empfehlung ist.

7. Verfahren nach Anspruch 1, wobei, falls eine Tonhöhenverschiebung (Pitch lag) des Unterrahmens weniger als ein vorbestimmter Parameter ist, die Anzahl der abgesuchten Verstärkungsskalierungsfaktoren von vier auf zwei reduziert ist.

8. Verfahren nach Anspruch 7, wobei, falls die Tonhöhenverschiebung des Unterrahmens gleich oder größer ist als der vorbestimmte Parameter, die Anzahl der abgesuchten Verstärkungsskalierungsfaktoren von vier auf drei reduziert ist.

9. Verfahren nach Anspruch 8, wobei der vorbestimmte Parameter 58 ist.

10. Verfahren nach Anspruch 1 oder 9, wobei die Quantisierungsstufen bzw. -Schritte und eine Vorsuche einmal durchgeführt werden, wenn die Tonhöhenverschiebung größer oder gleich 58 ist und zweimal ausgeführt wird, falls die Tonhöhenverschiebung weniger als 58 ist.

11. Ein Sprachkodiersystem mit zwei Raten bzw. Geschwindigkeiten, das eine reduzierte Rechenlast besitzt, wobei das Kodiersystem Mehrfachimpuls-Maximimal-Wahrscheinlichkeits-Quantisierungs-(Multi-Pulse Maximum Likelihood Quantization, MP-MLQ)-Verarbeitungsmittel besitzt zum Senden bzw. Übertragen mit einer ersten Übertragungsrate bzw. -geschwindigkeit und algebraische kodeangeregte Linearprediktions-(Algebraic Code-Excited Linear Prediction, ACELP)-Verarbeitungsmittel besitzt zum Senden bzw. Übertragen mit einer zweiten Übertragungsrate, wobei die normalen MP-MLQ-Verarbeitungsmittel-Unterrahmen von Anregungssignalen absucht und zwar gemäß einer nominalen Anzahl von Verstärkungsskalierungsfaktoren bei Quantisierung der Sprachsignale, und die normalen ACELP-Verarbeitungsmittel einen ersten Korrelationsschwellentest, zum Zulassen eines Eintritts in eine letzte Signalverarbeitungsschleife verwenden, dadurch gekennzeichnet dass:

die MP-MLQ-Verarbeitungsmittel eine reduzierte Anzahl von Verstärkungsskalierungsfaktoren zum Reduzieren der Anzahl von Verstärkungssuchen besitzen und dadurch die Rechenlast reduzieren;

die ACELP-Verarbeitungsmittel einen zweiten Korrelationsschwellentest verwenden, und zwar zum Zulassen eines Eintritts in eine vorhergehende Signalverarbeitungsschleife, die der genannten letzten Signalverarbeitungsschleife vorangeht in die in Abhängigkeit von dem genannten ersten Korrelationsschwellentest eingetreten wird, dadurch die Anzahl von Malen reduziert, die in die vorhergehende Signalverarbeitungsschleife und die letzte Signalverarbeitungsschleife eingetreten wird, was wiederum die Rechenlast reduziert.

Revendications

1. Procédé pour réduire la charge de calcul d'un système de codage à double vitesse, le système de codage étant configuré pour transmettre à une première vitesse de transmission en utilisant un processus de Quantification à Probabilité Maximale Multi-impulsion (MP-MLQ) ou à une seconde vitesse de transmission en utilisant un processus de Prédiction Linéaire Excitée en Code Algébrique (ACELP), dans lequel le processus MP-MLQ normal recherche des sous-trames de signaux d'excitation conformément à un nombre nominal de facteurs d'échelle de gain lors de l'exécution d'étapes de quantification pour coder les signaux vocaux et le processus ACELP normal impose un premier test de seuil de corrélation pour pénétrer une dernière boucle de traitement de signal, caractérisé en ce que le procédé inclut les étapes consistant à :

pour le processus MP-MLQ, réduire le nombre de facteurs d'échelle de gain utilisés lors des étapes de quantification, de manière à réduire le nombre de recherches de gain, ce qui réduit à son tour la charge de calcul, ou

pour le processus ACELP, imposer un second test de seuil de corrélation pour pénétrer une boucle précédente de traitement de signal qui précède ladite dernière boucle de traitement de signal pénétrée en fonction dudit premier test de seuil de corrélation, de manière à réduire le nombre de fois où la boucle précédente de traitement de signal et la dernière boucle de traitement de signal sont pénétrées, ce qui réduit à son tour la charge de calcul.

2. Procédé selon la revendication 1, dans lequel le second test de seuil est applicable pour pénétrer dans les troisième ou quatrième boucles de traitement de signal.

3. Procédé selon la revendication 2, dans lequel si la seconde vitesse de transmission est applicable, le procédé inclut en outre l'étape consistant à remplacer le premier seuil par un seuil plus élevé pour pénétrer dans la quatrième boucle de traitement de signal.

4. Procédé selon la revendication 3, incluant en outre, si la seconde vitesse de transmission est applicable, l'étape consistant à limiter le nombre de fois où la troisième ou quatrième boucle de traitement signal peut être pénétrée.

5. Procédé selon la revendication 4, dans lequel les troisième ou quatrième boucles de traitement de signal peuvent être pénétrées jusqu'à 32 ou 75 fois respectivement pour chacune des sous-trames vocales.

6. Procédé selon la revendication 5, dans lequel le système de codage à double vitesse est généralement conforme à la recommandation ITU-T G.723.1.

7. Procédé selon la revendication 1, dans lequel si un décalage de pas de la sous-trame est inférieur à un paramètre prédéterminé, le nombre de facteurs d'échelle de gain recherchés est réduit de quatre à deux.

8. Procédé selon la revendication 7, dans lequel si le décalage de pas de la sous-trame est égal ou supérieur au paramètre prédéterminé, le nombre de facteurs d'échelle de gain recherchés est réduit de quatre à trois.

9. Procédé selon la revendication 8, dans lequel le paramètre prédéterminé est égal à 58.

10. Procédé selon la revendication 1 ou 9, dans lequel les étapes de quantification et une prérecherche sont exécutées une fois si le décalage de pas est supérieur ou égal à 58 et deux fois si le décalage de pas est inférieur à 58.

11. Système de codage vocal à double vitesse ayant une charge de calcul réduite, le système de codage ayant des moyens de traitement par Quantification à Probabilité Maximale Multi-impulsion (MP-MLQ) pour transmettre à une première vitesse de transmission et des moyens de traitement par Prédiction Linéaire Excitée en Code Algébrique (ACELP) pour transmettre à une seconde vitesse de transmission, dans lequel les moyens de traitement MP-MLQ normaux recherchent des sous-trames de signaux d'excitation conformément à un nombre nominal de facteurs d'échelle de gain lors de la quantification des signaux vocaux, et les moyens de traitement ACELP normaux utilisent un premier test de seuil de corrélation pour permettre de pénétrer dans une dernière boucle de traitement de signal, caractérisé en ce que :

les moyens de traitement MP-MLQ ont un nombre réduit de facteurs d'échelle de gain pour réduire le nombre de recherches de gain et réduire ainsi la charge de calcul,

les moyens de traitement ACELP utilisent un second test de seuil de corrélation pour permettre de pénétrer dans une boucle précédente de traitement de signal qui précède ladite dernière boucle de traitement de signal pénétrée en fonction dudit premier test de seuil de corrélation, de manière à réduire le nombre de fois où la boucle précédente de traitement de signal et la dernière boucle de traitement de signal sont pénétrées, ce qui réduit à son tour la charge de calcul.

Drawing